Artificial Intelligence on the Edge — Brisbane, Australia
We build free, open-source Android apps that run AI entirely on your device. No cloud. No subscription. No data leaving your hands.
Cloud AI is powerful, but it comes with trade-offs: latency, cost, and your data leaving your device. We build tools that bring modern AI models to mobile hardware, so your conversations and voice stay entirely yours — no server, no compromise.
- Private — no telemetry, no ads, no tracking
- Offline — works anywhere, no internet required
- Open Source — MIT licensed, community driven
Maid — Mobile Artificial Intelligence Distribution
A fully-featured Android AI chat app. Run GGUF models locally via llama.cpp, or connect to remote providers using your own API key.
- Local inference via llama.cpp — fully on-device, no internet needed
- Remote providers — Anthropic, DeepSeek, Mistral, Ollama, OpenAI
- One-tap model downloads from Hugging Face
- Conversation management — create, rename, export, and import chats
- Customisable generation parameters and system prompts
- Voice output via Maise integration
- Optional account sync via Supabase
- Material You theming — light and dark
Maise — Mobile Artificial Intelligence Speech Engine
An on-device Android text-to-speech engine with 40+ voices across 9 languages, powered by ONNX Runtime.
- Implements the standard Android
TextToSpeechAPI — works system-wide with any app - Two neural TTS engines: Kokoro and Kitten TTS
- Languages: English (US/UK), German, French, Greek, Italian, Japanese, Portuguese, Chinese
- Streaming playback — audio starts before synthesis completes
- 24 kHz mono 16-bit PCM output